Intercellular mechanisms by which the stromal microenvironment contributes to solid tumor progression and targeted therapy resistance remain poorly understood, presenting significant clinical hurdles. PEAK1 (Pseudopodium-Enriched Atypical Kinase One) is an actin cytoskeleton- and focal adhesion-associated pseudokinase that promotes cell state plasticity and cancer metastasis by mediating growth factor-integrin signaling crosstalk. Here, we determined that stromal PEAK1 expression predicts poor outcomes in HER2-positive breast cancers high in SNAI2 expression and enriched for MSC content. Specifically, we identified that the fibroblastic stroma in HER2-positive breast cancer patient tissue stains positive for both nuclear SNAI2 and cytoplasmic PEAK1. Furthermore, mesenchymal stem cells (MSCs) and cancer-associated fibroblasts (CAFs) express high PEAK1 protein levels and potentiate tumorigenesis, lapatinib resistance and metastasis of HER2-positive breast cancer cells in a PEAK1-dependent manner. Analysis of PEAK1-dependent secreted factors from MSCs revealed INHBA/activin-A as a necessary factor in the conditioned media of PEAK1-expressing MSCs that promotes lapatinib resistance. Single-cell CycIF analysis of MSC-breast cancer cell co-cultures identified enrichment of p-Akthigh/p-gH2AXlow, MCL1high/p-gH2AXlow and GRP78high/VIMhigh breast cancer cell subpopulations by the presence of PEAK1-expressing MSCs and lapatinib treatment. Bioinformatic analyses on a PEAK1-centric stroma-tumor cell gene set and follow-up immunostaining of co-cultures predict targeting antiapoptotic and stress pathways as a means to improve targeted therapy responses and patient outcomes in HER2-positive breast cancer and other stroma-rich malignancies. These data provide the first evidence that PEAK1 promotes tumorigenic phenotypes through a previously unrecognized SNAI2-PEAK1-INHBA stromal cell axis.

Cancer metastasis is responsible for the clear majority of cancer-related deaths. Survival and expansion of cancer cells at secondary sites requires that these premetastatic microenvironments be primed by primary tumor cells and their secreted factors. Efforts to date have been limited by immune-deficient in vivo models and/or the need for finely-tuned analysis time points that reduce contributions from early-disseminating cancer cells. In this regard, we developed a tumor cell-free syngeneic breast cancer model for characterizing tumor cell secretome-mediated reprogramming of premetastatic tissues. We demonstrate that secretomes from metastatic breast cancer cells differentially regulate the lung and brain, promoting a tumor-supportive lung microenvironment with both elevated CD73 expression and decreased TNFα expression. Using in vitro models of CD73-positive mesenchymal stem cells (MSCs) and macrophages/monocytes, we tested whether MSCs can mediate anti-inflammatory effects of metastatic breast cancer cells. Notably, conditioned media from metastatic Py230 cells reprogrammed the secretomes of MSCs toward an anti-inflammatory state. Mining transcriptome data from Py8119 and Py230 cells revealed a lipocalin 2 (LCN2) axis that is selectively expressed in the metastatic Py230 cells, predicts poor breast cancer patient survival and is elevated in circulating serum of mice chronically treated with conditioned media from Py230 cells. Taken together, these results establish the utility of an immune-competent tumor cell-free model for characterizing the mechanisms of breast cancer cell priming of the premetastatic niche, demonstrate that MSCs can mediate the anti-inflammatory effects of metastatic breast cancer cells and substantiate LCN2 as a promising therapeutic target for blocking breast cancer progression.

No abstract